Efficient water quality prediction models based on machine learning algorithms for Nainital Lake, Uttarakhand. (2022)

• Record Type: Journal Article
• Authors: Koranga, Manisha; Pant, Pushpa; Kumar, Tarun; Pant, Durgesh; Bhatt, Ashutosh Kumar; Pant, R.P.
• Is Part Of: Materials today. Volume 57:Part 4(2022)
• Page Start: 1706
• Page End: 1712
• Publication Date: 2022
• Subjects: Nainital Lake -- Water quality -- Machine learning algorithms -- Classification -- Regression; Materials science -- Congresses -- Periodicals; 620.1
• DOI: 10.1016/j.matpr.2021.12.334
• Languages: English
• ISSNs: 2214-7853

Abstract

Abstract: Water quality deterioration increases day by day in hilly areas due to increasing tourism activity, unplanned construction, disposal of solid waste, improper sewage management. With this idea, the work investigates different machine learning algorithms to evaluate the water quality index (WQI) and the water quality class (WQC). This paper utilizes Nainital Lake as a study area. The models used for testing and training comprise algorithms of machine learning for both binary and multiclass classification. In this paper, eight machine learning algorithms were employed for regression analysis, and nine machine learning algorithms were used for classification analysis. The result demonstrates that in regression analysis, the Random Forest algorithm comes out to be the most efficient Machine Learning algorithm. However, in the case of classification analysis, no single algorithm is good enough for prediction, three algorithms Stochastic Gradient Descent, Random Forest, and Support Vector Machine with the same accuracy proved to be efficient to predict water quality.